| 格网划分的最邻近点搜索方法 |
| |
| 引用本文： | 杨容浩,岑敏仪,张同刚,郑平元.格网划分的最邻近点搜索方法[J].测绘科学,2012,37(5):90-93. |
| |
| 作者姓名： | 杨容浩 岑敏仪 张同刚 郑平元 |
| |
| 作者单位： | 1. 西南交通大学地球科学与环境工程学院,成都610031;成都理工大学地球科学学院,成都610051
2. 西南交通大学地球科学与环境工程学院,成都,610031 |
| |
| 摘 要： | 为了提高迭代最近点(ICP)算法中最邻近点搜索的存储和计算效率,本文通过对盒子结构方法优、缺点的深入分析,提出了基于格网划分的最邻近点搜索方法。该方法充分考虑了3D点云获取时的投影特性,将点云投影到某一坐标平面,并基于格网划分进行存储,使最邻近点的搜索限制在较小的范围。不同类型的模拟数据和实测数据试验均表明,该方法能够在不损失匹配精度和拉入范围的前提下,显著提高存储和计算效率。
|
| 关 键 词： | D点云 表面匹配 迭代最近点算法(ICP) 最邻近点搜索 盒子结构 格网划分 |
A method of closest points searching based on grid partition |
| |
| YANG Rong-hao
,
CEN Min-yi
,
ZHANG Tong-gang
,
ZHENG Ping-yuan.A method of closest points searching based on grid partition[J].Science of Surveying and Mapping,2012,37(5):90-93. |
| |
| Authors: | YANG Rong-hao CEN Min-yi ZHANG Tong-gang ZHENG Ping-yuan |
| |
| Institution: | ①(①Faculty of Geosciences and Environmental Engineering,Southwest Jiaotong University,Chengdu 610031,China;②College of Earth Sciences,Chengdu University of Technology,Chengdu 610051,China) |
| |
| Abstract: | In order to improve the storage and computational efficiency of closest points searching of the Iterative Closest Point(ICP) algorithm,by a deep analysis of advantages and disadvantages about the method of boxing structure,a method of closest points searching based on grid partition was proposed in the paper.This method considered the projection characteristic of the 3D points cloud being acquired fully,and projected the points cloud to one of the coordinate planes,and stored the points cloud based on grid partition,then limited the closest points searching in a smaller extent.Both the tests with simulated data of different kinds and real measured data showed that this method could significantly improve the storage and computational efficiency without losing the registration precision and pull-in range. |
| |
| Keywords: | 3D points cloud surface registration Iterative Closest Point(ICP) algorithm closest points searching boxing structure grid partition |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
|
